The major objective of this grant application is to develop methods to detect AIDS in blood samples for transfusion, the potential to transmit AIDS, so that such blood specimens can be eliminated from the blood donor pool. The study population will consist of 350 male homosexuals of whom 20% are symptom free, 60% have the AIDS-Related Symptom Complex (ARC), and 20% have AIDS with Kaposi's sarcoma, Opportunistic Infection, or both. In addition, 500 (or more if necessary) blood bank donor specimens, 100 primary breast cancer patients, and 100 acute leukemia patients will be studied. The latter patient group will be evaluated at diagnosis and one year later when they usually have received multiple blood transfusions and immunosuppressive chemotherapy. Also, 100 female hospital employees with no AIDS contacts will be additional controls. The premise of the study will be that one single test may either over-\or underestimate the potential to transmit AIDS. Therefore, a battery of nine tests will be applied to 15 ml of blood from each donor and we will select two or three parametera which identify 100% of ARC patients with lymphadenopathy, 100% of patients who move from symptom free to ARC or from ARC to AIDS, and 100% of AIDS patients (with the precaution that some of the potentially positive parameters may become negative in patients with advanced disease). The parameters to be studied will include CMV antibody titer (indirect immunofluorescence), antithymic antibody titer (indirect immunofluorescence), immune complexes (Clq biding assay, modified for ELISA), serum alpha-interferon titers (including acid lability by cytopathic effect modification), serum beta-2 microglobulin level (radioimmunoassay), blood levels of cella marking for the T3, T4, and T8 antigens, with calculation of the 4:8 ratio and lymphocytes (flow cytometry), 5'-nucleotidase levels in terms of enzyme activity /10[unreadable]6[unreadable] /hr (for conversion of C[unreadable]14[unreadable]-AMP substrate). Data will be analyzed by chi-squared and ANOVA tests, and multivariant logistic regression analysis. Evaluation of 500 normal blood bank specimens will allow an estimate of cost effectiveness procedures. Evaluation of transfused leukemia patients will give estimate of clinical risk. (SR)